---
title: 'Air Quality Prediction'
output: 
  flexdashboard::flex_dashboard:
    orientation: columns
    vertical_layout: fill
    theme: flatly
    social: [ "twitter", "facebook", "menu" ]
    source_code: embed
        

---

```{r setup, include=FALSE}
library(flexdashboard)
library(tidyverse)
library(htmltab)
library(tsbox)
library(stR)
library(quantmod)
library(tseries)
library(forecast)
library(dygraphs)
library(leaflet)
library(ggplot2)
library(plotly)
library(readxl)


data <- read.csv('~/R/AQI prediction/data2.csv')
data$Date<-as.Date(data$Date)
AQI <- read.csv('~/R/AQI prediction/Data/city_day.csv')


AQID<-dplyr::filter(AQI,City=="Delhi")
AQIM<-dplyr::filter(AQI,City=="Mumbai")
AQIC<-dplyr::filter(AQI,City=="Chennai")

AQID<-AQID[,c("Date","AQI")]
AQID$Date<-as.Date(AQID$Date,format="%d-%m-%Y")
AQIM<-AQIM[,c("Date","AQI")]
AQIM$Date<-as.Date(AQIM$Date,format="%d-%m-%Y")
AQIC<-AQIC[,c("Date","AQI")]
AQIC$Date<-as.Date(AQIC$Date,format="%d-%m-%Y")

AQID.ts<-ts(AQID$AQI,start=c(2015,1),frequency=365)
AQID.ts<-na.approx(AQID.ts)
AQIM.ts<-ts(AQIM$AQI,start=c(2015,1),frequency=365)
AQIM.ts<-na.approx(AQIM.ts)
AQIC.ts<-ts(AQIC$AQI,start=c(2015,1),frequency=365)
AQIC.ts<-na.approx(AQIC.ts)

```



```{r , include=FALSE}
worldPM1025 <- read_excel("aap_air_quality_database_2018_v14.xlsx",  sheet = "database", 
                          col_types = c("text","skip", "text", "skip", "skip", "text", "skip", "skip", 
                                        "text", "skip", "skip", "skip", "skip", "skip", "skip"), skip = 2)

#Source: WHO Global Ambient Air Quality Database - update 2018

#Preprocessing
names(worldPM1025)<-c("Region", "Country","PM10","PM2.5")
worldPM1025$PM10<-stringr::str_replace(worldPM1025$PM10,"-converted value","")
worldPM1025$PM10<-stringr::str_replace(worldPM1025$PM10,"[)]","")
worldPM1025$PM10<-stringr::str_replace(worldPM1025$PM10,"[(]","")
worldPM1025$PM2.5<-stringr::str_replace(worldPM1025$PM2.5,"-converted value","")
worldPM1025$PM2.5<-stringr::str_replace(worldPM1025$PM2.5,"[)]","")
worldPM1025$PM2.5<-stringr::str_replace(worldPM1025$PM2.5,"[(]","")
worldPM1025$PM2.5<-as.numeric(worldPM1025$PM2.5)
worldPM1025$PM10<-as.numeric(worldPM1025$PM10)
World2.5<-aggregate(x=worldPM1025$PM2.5,by=list(Country=worldPM1025$Country,Region=worldPM1025$Region),FUN = mean)
World10<-aggregate(x=worldPM1025$PM10,by=list(Country=worldPM1025$Country,Region=worldPM1025$Region),FUN = mean)
          

#Source:https://www.kaggle.com/paultimothymooney/latitude-and-longitude-for-every-country-and-state
countries <- read.csv("~/countries.csv")
names(countries)<-c("code","lat","lon","Country")

mergeddata2.5<-merge(World2.5,countries,type="left")
mergeddata10<-merge(World10,countries,type="left")

# Create a palette that maps factor levels to colors
pal <- colorNumeric(palette = "YlOrRd", domain = mergeddata2.5$x)
pal2 <- colorNumeric(palette = "YlOrRd", domain = mergeddata10$x)

```


```{r, include = FALSE}

o=data.frame(na.omit(AQI[,c(1,5:9)]))
fin<-aggregate(o,by=list(o$City),FUN=mean)
fin<-na.omit(fin[,-2])
names(fin)=c("city","NO","NO2","NOx","NH3","CO")


indlatlon <- read.csv("~/in.csv")

indiaf<-merge(fin,indlatlon,type="inner",by="city")

```


Overview {data-navmenu="Intro"}
===================================== 


Column
-------------------------------------
    
### Overview

The Problem


Climate change is a very big problem in contemporary times, and India, as the second most populous country in the world, contributes significantly to air pollution with many cities in the top 50 most polluted cities in the world. But the extent of pollution is not properly studied and used in decision making in India. Air quality monitoring stations are present in many cities but the data isn’t used for predictions in most cases, but rather as an observation. Another problem is that these AQI monitoring stations aren’t present in small cities. The project aims to build a forecasting model to predict AQI. The project focuses on the city of Indian cities like Delhi, Mumbai and Chennai which are infamous for their pollution levels.





### India NOx pollution

```{r}

indiafplot3<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$NOx)
pal5 <- colorNumeric(palette = "YlOrRd", domain = indiafplot3$x)
leaflet(indiafplot3) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x,
    color = ~pal5(x),
    stroke = FALSE, fillOpacity = 0.5
  )
```
   
Column
-------------------------------------
   
### Pollution level by continents and income (PM 10)

```{r}
library(plotly)
fig <- plot_ly(World10, y = ~x, color = ~Region, type = "box")

fig
```   
 
### Polution levels in different countries (PM 10)
    
```{r}

leaflet(mergeddata10) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x/5,
    color = ~pal2(x),
    stroke = FALSE, fillOpacity = 0.5
  )
```



Data Summary {data-navmenu="Intro"}
===================================== 

Column
-------------------------------------

### Meteorological data
Meteorological data


The Meteorological data was scrapped from tutiempo.com. The coding for the scrapper is given with the file.The values collected were Temperature, Purssure, Humidity and Wind Speed.

### AQI Data
AQI Data


The AQI data was downloaded from Kaggle. The data set has AQI, CO, NO, NOx, NH3, SO2, PM2.5, PM10 emissions measured in micro grams per meter cube.

### World pollution Data
World pollution Data


This data was downloaded from WHO Global Ambient Air Quality Database - update 2018.It had locationwise pollution data for countries

Column
-------------------------------------

### India pollution Data
India pollution Data


The AQI data was manipulated and grouped to obtain mean pollution across various cities in India.

### Latitude Longitude countries
Latitude Longitude countries


This Data was downloaded from Kaggle.Since goggle's ggmap is no more free. This was a better option than buying an API.

### Latitude Longitude Indian cities
Latitude Longitude Indian cities


This data had location (geocode) of different Indian cities. This is matched with existing pollution data for visualizations.

World Pollution Levels {data-navmenu="EDA"}
=========================================

Column
-------------------------------------

### Plot Inference 

Pollution Levels around the world

The plots on the side reveal how developing countries (Low Income Mediterranean, South East Asia) contribute most to the particulate pollution. Moreover the other plots reveal how India compares to the rest of the world
LMIC: Low Middle Income Countries
HIC: High Income Countries

### PM 2.5 across countries mapped
```{r}
leaflet(mergeddata2.5) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x/5,
    color = ~pal(x),
    stroke = FALSE, fillOpacity = 0.5
  )

```




### PM 10 across countries mapped
```{r}
leaflet(mergeddata10) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x/5,
    color = ~pal2(x),
    stroke = FALSE, fillOpacity = 0.5
  )

```


Column
------------------------------------- 


### PM 10 across continents
```{r}

library(plotly)
fig <- plot_ly(World10, y = ~x, color = ~Region, type = "box")

fig
```




### PM 2.5 across continents 
```{r}
library(plotly)
fig1 <- plot_ly(World2.5, y = ~x, color = ~Region, type = "box")

fig1
```
 


India Pollution Levels{ data-navmenu="EDA"}
=========================================

Column
------------------------------------- 

### Plot Inference 
Pollution Levels in India

The plots on the side show the amounts of pollutants and their spread across various cities in India. Particulate Matter is not shown here. The pollutants shown here make up significant portion of the Air Quality Index which is a widely used measure. We will be using the AQI as well in this project




### CO

```{r}


indiafplot1<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$CO)

pal3 <- colorNumeric(palette = "YlOrRd", domain = indiafplot1$x)
leaflet(indiafplot1) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x*15,
    color = ~pal3(x),
    stroke = FALSE, fillOpacity = 0.5
  )




```




### NO2

```{r}

indiafplot2<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$NO2)
pal4 <- colorNumeric(palette = "YlOrRd", domain = indiafplot2$x)
leaflet(indiafplot2) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x,
    color = ~pal4(x),
    stroke = FALSE, fillOpacity = 0.5
  )






```

Column
------------------------------------- 

### NOx

```{r}

indiafplot3<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$NOx)
pal5 <- colorNumeric(palette = "YlOrRd", domain = indiafplot3$x)
leaflet(indiafplot3) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x,
    color = ~pal5(x),
    stroke = FALSE, fillOpacity = 0.5
  )



```

 

### NO
```{r}

indiafplot4<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$NO)
pal6 <- colorNumeric(palette = "YlOrRd", domain = indiafplot4$x)
leaflet(indiafplot4) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x/2,
    color = ~pal6(x),
    stroke = FALSE, fillOpacity = 0.5
  )


```

### NH3
```{r}

indiafplot5<-data.frame(lat=indiaf$lat,lon=indiaf$lng,x=indiaf$NH3)
pal7 <- colorNumeric(palette = "YlOrRd", domain = indiafplot5$x)
leaflet(indiafplot5) %>% addTiles() %>%
  addCircleMarkers(
    radius = ~x/2,
    color = ~pal7(x),
    stroke = FALSE, fillOpacity = 0.5
  )
```

Delhi{data-navmenu="Meterological Trends"}
=========================================

Column
------------------------------------- 

### Delhi
```{r}
m <- leaflet() %>%
  addTiles() %>% 
  addMarkers(lng=77.1025, lat=28.7041, popup="Delhi")
m 
```


### Temperature


```{r}
datadelhi <- read.csv("datadelhi.csv")
datadelhi$Date<-as.Date(datadelhi$Date)
x <- datadelhi$Date
y<-datadelhi$temp
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Degrees"))

fig
```

Column
------------------------------------- 

### Pressure

```{r}
x <- datadelhi$Date
y<-datadelhi$pressure
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("hPa"))

fig
```



### Humidity

```{r}
x <- datadelhi$Date
y<-datadelhi$humidity[datadelhi$humidity<1000]
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("%"))

fig
```

### Wind

```{r}
x <- datadelhi$Date
y<-datadelhi$wind
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Km/h"))

fig
```


Mumbai{data-navmenu="Meterological Trends"}
=========================================


Column
-------------------------------------


### Mumbai

```{r}
m <- leaflet() %>%
  addTiles() %>% 
  addMarkers(lng=72.8777, lat=19.0760, popup="Mumbai")
m 
```


### Temperature


```{r}
datamumbai <- read.csv("datamumbai.csv")
datamumbai$Date<-as.Date(datamumbai$Date)
x <- datamumbai$Date
y<-datamumbai$temp
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Degrees"))

fig
```

Column
-------------------------------------

### Pressure

```{r}
x <- datamumbai$Date[datamumbai$pressure<1100]
y<-datamumbai$pressure[datamumbai$pressure<1100]
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("hPa"))

fig
```



### Humidity

```{r}
x <- datamumbai$Date
y<-datamumbai$humidity
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("%"))

fig
```

### Wind

```{r}
x <- datamumbai$Date
y<-datamumbai$wind
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Km/h"))

fig
```


Chennai{data-navmenu="Meterological Trends"}
=========================================

Column
-------------------------------------

### Chennai

```{r}
m <- leaflet() %>%
  addTiles() %>% 
  addMarkers(lng=80.2707, lat=13.0827, popup="Chennai")
m 
```


### Temperature


```{r}

datachennai <- read.csv("datachennai.csv")
datachennai$Date<-as.Date(datachennai$Date)
x <- datachennai$Date
y<-datachennai$temp
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Degrees"))

fig
```


Column
-------------------------------------

### Pressure

```{r}
x <- datachennai$Date
y<-datachennai$pressure
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("hPa"))

fig
```

### Humidity

```{r}
x <- datachennai$Date
y<-datachennai$humidity
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("%"))

fig
```

### Wind

```{r}
x <- datachennai$Date[datachennai$wind<22 ]
y<-datachennai$wind[datachennai$wind<22 ]
fig <- plot_ly(x = ~x, y = ~y, mode = 'lines', text = paste("Km/h"))

fig
```



Methodology {data-navmenu="Time Series Analysis"}
=========================================

Column
-------------------------------------

### ACF/ PACF
Meteorological data


By looking at the autocorrelation function (ACF) and partial autocorrelation (PACF) plots we can tentatively identify the numbers of AR and/or MA terms that are needed.ACF plot is merely a bar chart of the coefficients of correlation between a time series and lags of itself. The PACF plot is a plot of the partial correlation coefficients between the series and lags of itself.

### Model Fitting
AQI Data


We have used the algorithm proposed by Hyndman & Khandakarin 2008 which is implemented through auto.arima in R. This simplifies the process but increases time complexity. Since we have a relatively small data set we have gone ahead with this.

### SARIMA
SARIMA


SARIMA is a Integrated AR, MA model fitted for daily/monthly values as well as an additional seasonal term. We have fitted the model using auto.arima.

Column
-------------------------------------

### SARIMAX
SARIMAX


The SARIMAX model has some regressors as well, to model the variable change with time. This model was fitted using auto.arima but the results were unfavorable due to missing values and compounding of errors.A Distributed Lag model was also checked but gave worse results.

### Ljung Box Test
Ljung Box Test


The Ljung Box test of the residuals show if the residuals are white noise or if they have predictability. The squared residual plot checks conditional heteroscedasticity.

### Model Selection
Model Selections


There are various error measures that can be used for model selection - AIC, BIC, MSE, MAE, MAPE, etc. We have used MAPE. We have chosen the most robust model as that is also an important criteria in selection.


ACF/PACF {data-navmenu="Time Series Analysis"}
=========================================

### Importance {.tabset}
-------------------------------------

### ACF Delhi
```{r}

conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQID.ts))
bacf <- pacf(AQID.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")
q

```

### PACF Delhi
```{r}



conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQID.ts))
bacf <- acf(AQID.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")
q
```


### ACF Mumbai
```{r}

conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQIM.ts))
bacf <- pacf(AQIM.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")

q
```

### PACF Mumbai
```{r}



conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQIM.ts))
bacf <- acf(AQIM.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")
q
```



### ACF Chennai
```{r}

conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQIC.ts))
bacf <- pacf(AQIC.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")

q
```

### PACF Chennai
```{r}



conf.level <- 0.95
ciline <- qnorm((1 - conf.level)/2)/sqrt(length(AQIC.ts))
bacf <- acf(AQIC.ts, plot = FALSE)
bacfdf <- with(bacf, data.frame(lag, acf))

library(ggplot2)
q <- ggplot(data=bacfdf, mapping=aes(x=lag, y=acf)) +
       geom_bar(stat = "identity", position = "identity")
q
```

SARIMA {data-navmenu="Time Series Analysis"}
=========================================

Column
-------------------------------------
### Summary
Fitted Models
Delhi: SARIMA(1,1,2)(0,1,0)
Mumbai: SARIMA(1,0,0)(0,1,0)
Chennai: SARIMA(4,1,3)
The Chennai model has a lot of missing values and hence cannot be predicted properly, hence the apparent lack of seasonality. We drop Chennai from our analysis.


### Delhi
```{r}

set.seed(1000)
ARIMAXD.out=auto.arima(y = AQID.ts)

forcastD<-forecast(object = ARIMAXD.out,h = 365)
Time0 <-time(forcastD$mean)
AQIDelhi0<-as.numeric(forcastD$mean)
figD <- plot_ly(x = ~Time0, y = ~AQIDelhi0, mode = 'lines',type = "scatter")

#forcastD

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastD$fitted),
  y = as.numeric(forcastD$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastD$upper),(append((rev(time(forcastD$upper))[1]),rev(time(forcastD$upper))))),
  y = append(as.numeric(forcastD$lower[,2]),append((rev(as.numeric(forcastD$upper[,2]))[1]),rev(as.numeric(forcastD$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastD$upper),(append((rev(time(forcastD$upper))[1]),rev(time(forcastD$upper))))),
  y = append(as.numeric(forcastD$lower[,1]),append((rev(as.numeric(forcastD$upper[,1]))[1]),rev(as.numeric(forcastD$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastD$mean),
  y = as.numeric(forcastD$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR", 
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p




```

Column
-------------------------------------

### Mumbai
```{r}


set.seed(1000)
ARIMAXM.out=auto.arima(y = AQIM.ts)


forcastM<-forecast(object = ARIMAXM.out,h = 365)


Time1<-time(forcastM$mean)
AQIMumbai0<-as.numeric(forcastM$mean)
figM <- plot_ly(x = ~Time1, y = ~AQIMumbai0, mode = 'lines',type = "scatter")



#forcastM

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastM$fitted),
  y = as.numeric(forcastM$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastM$upper),(append((rev(time(forcastM$upper))[1]),rev(time(forcastM$upper))))),
  y = append(as.numeric(forcastM$lower[,2]),append((rev(as.numeric(forcastM$upper[,2]))[1]),rev(as.numeric(forcastM$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastM$upper),(append((rev(time(forcastM$upper))[1]),rev(time(forcastM$upper))))),
  y = append(as.numeric(forcastM$lower[,1]),append((rev(as.numeric(forcastM$upper[,1]))[1]),rev(as.numeric(forcastM$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastM$mean),
  y = as.numeric(forcastM$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR", 
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p


```

### Chennai
```{r}

set.seed(1000)
ARIMAXC.out=auto.arima(y = AQIC.ts)


forcastC<-forecast(object = ARIMAXC.out,h = 365)


x1 <-seq( from =as.Date("2020-01-08"),to = as.Date("2021-01-08"),length.out = 365)
y1<-as.numeric(forcastC$mean)
fig <- plot_ly(x = ~x1, y = ~y1, mode = 'lines',type = "scatter")

#forcastC

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastC$fitted),
  y = as.numeric(forcastC$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastC$upper),(append((rev(time(forcastC$upper))[1]),rev(time(forcastC$upper))))),
  y = append(as.numeric(forcastC$lower[,2]),append((rev(as.numeric(forcastC$upper[,2]))[1]),rev(as.numeric(forcastC$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastC$upper),(append((rev(time(forcastC$upper))[1]),rev(time(forcastC$upper))))),
  y = append(as.numeric(forcastC$lower[,1]),append((rev(as.numeric(forcastC$upper[,1]))[1]),rev(as.numeric(forcastC$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastC$mean),
  y = as.numeric(forcastC$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR", 
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p

```

Model Results SARIMA { data-navmenu="Time Series Analysis"}
===================================================================


### Importance {.tabset}
-------------------------------------

### Delhi
```{r}
print("Delhi")
summary(ARIMAXD.out)

```

### Mumbai

```{r}
print("Mumbai")
summary(ARIMAXM.out)
```

### Chennai

```{r}

print("Chennai")
summary(ARIMAXC.out)

```


SARIMAX {data-navmenu="Time Series Analysis"}
=========================================

Column
-------------------------------------

### Summary
Fitted Models
Delhi: SARIMAX(1,0,0)
Mumbai: SARIMA(2,1,2)
Chennai: SARIMA(1,1,1)
These ARIMAX models are done using all the meteorological variables. This gives models with larger MAPE than the ARIMA model without an X variable/ matrix. This can be due to high deviations in the meteorological variables. A distributed lagged model was also fitted to check. ARIMA with no X variable is the best in any case.


### Delhi
```{r}


datadelhi <- read.csv("datadelhi.csv")
datadelhi$Date<-as.Date(datadelhi$Date)
ydl=as.ts(AQIM.ts[1:1500])
xdl=as.matrix(datadelhi[1:1500,-1])
ARIMAXD2.out=auto.arima(y = ydl, xreg = xdl)


forcastdlD<-forecast(object = ARIMAXD2.out,h = 365,xreg=as.matrix(datadelhi[1501:1865,-1]))

#forcastdlD

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastdlD$fitted),
  y = as.numeric(forcastdlD$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastdlD$upper),(append((rev(time(forcastdlD$upper))[1]),rev(time(forcastdlD$upper))))),
  y = append(as.numeric(forcastdlD$lower[,2]),append((rev(as.numeric(forcastdlD$upper[,2]))[1]),rev(as.numeric(forcastdlD$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastdlD$upper),(append((rev(time(forcastdlD$upper))[1]),rev(time(forcastdlD$upper))))),
  y = append(as.numeric(forcastdlD$lower[,1]),append((rev(as.numeric(forcastdlD$upper[,1]))[1]),rev(as.numeric(forcastdlD$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastdlD$mean),
  y = as.numeric(forcastdlD$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR", 
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p






```

Column
-------------------------------------

### Mumbai
```{r}


datamumbai <- read.csv("datamumbai.csv")
datamumbai$Date<-as.Date(datamumbai$Date)
ydl=as.ts(AQIM.ts[1:1500])
xdl=as.matrix(datamumbai[1:1500,-1])
ARIMAXM2.out=auto.arima(y = ydl, xreg = xdl)


forcastdlM<-forecast(object = ARIMAXM2.out,h = 365,xreg=as.matrix(datamumbai[1501:1865,-1]))


#forcastdlM

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastdlM$fitted),
  y = as.numeric(forcastdlM$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastdlM$upper),(append((rev(time(forcastdlM$upper))[1]),rev(time(forcastdlM$upper))))),
  y = append(as.numeric(forcastdlM$lower[,2]),append((rev(as.numeric(forcastdlM$upper[,2]))[1]),rev(as.numeric(forcastdlM$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastdlM$upper),(append((rev(time(forcastdlM$upper))[1]),rev(time(forcastdlM$upper))))),
  y = append(as.numeric(forcastdlM$lower[,1]),append((rev(as.numeric(forcastdlM$upper[,1]))[1]),rev(as.numeric(forcastdlM$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastdlM$mean),
  y = as.numeric(forcastdlM$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR", 
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p




```

### Chennai
```{r}

datachennai <- read.csv("datachennai.csv")
datachennai$Date<-as.Date(datachennai$Date)
ydl=as.ts(AQIC.ts[1:1500])
xdl=as.matrix(datachennai[1:1500,-1])
ARIMAXC2.out=auto.arima(y = ydl, xreg = xdl)



forcastdlC<-forecast(object = ARIMAXC2.out,h = 365,xreg=as.matrix(datachennai[1501:1865,-1]))





#forcastdlC

library(plotly)
trace1 <- list(
  line = list(
    color = "rgba(0,0,0,1)", 
    fillcolor = "rgba(0,0,0,1)"
  ), 
  mode = "lines", 
  name = "observed", 
  type = "scatter", 
  x = time(forcastdlC$fitted),
  y = as.numeric(forcastdlC$fitted), 
  xaxis = "x", 
  yaxis = "y"
)
trace2 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(242,242,242,1)", 
    fillcolor = "rgba(242,242,242,1)"
  ), 
  mode = "lines", 
  name = "95% confidence", 
  type = "scatter", 
  x = append(time(forcastdlC$upper),(append((rev(time(forcastdlC$upper))[1]),rev(time(forcastdlC$upper))))),
  y = append(as.numeric(forcastdlC$lower[,2]),append((rev(as.numeric(forcastdlC$upper[,2]))[1]),rev(as.numeric(forcastdlC$upper[,2])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace3 <- list(
  fill = "toself", 
  line = list(
    color = "rgba(204,204,204,1)", 
    fillcolor = "rgba(204,204,204,1)"
  ), 
  mode = "lines", 
  name = "80% confidence", 
  type = "scatter", 
  x = append(time(forcastdlC$upper),(append((rev(time(forcastdlC$upper))[1]),rev(time(forcastdlC$upper))))),
  y = append(as.numeric(forcastdlC$lower[,1]),append((rev(as.numeric(forcastdlC$upper[,1]))[1]),rev(as.numeric(forcastdlC$upper[,1])))),
  xaxis = "x", 
  yaxis = "y", 
  hoveron = "points"
)
trace4 <- list(
  line = list(
    color = "rgba(0,0,255,1)", 
    fillcolor = "rgba(0,0,255,1)"
  ), 
  mode = "lines", 
  name = "prediction", 
  type = "scatter", 
  x = time(forcastdlC$mean),
  y = as.numeric(forcastdlC$mean),
  xaxis = "x", 
  yaxis = "y"
)
data <- list(trace1, trace2, trace3, trace4)
layout <- list(
  title = "Forecast from Auto.Arima", 
  xaxis = list(
    title = "Year", 
    domain = c(0, 1)
  ), 
  yaxis = list(
    title = "SAR",
    domain = c(0, 1)
  ), 
  margin = list(
    b = 40, 
    l = 60, 
    r = 10, 
    t = 25
  )
)
p <- plot_ly()
p <- add_trace(p, line=trace1$line, mode=trace1$mode, name=trace1$name, type=trace1$type, x=trace1$x, y=trace1$y, xaxis=trace1$xaxis, yaxis=trace1$yaxis)
p <- add_trace(p, fill=trace2$fill, line=trace2$line, mode=trace2$mode, name=trace2$name, type=trace2$type, x=trace2$x, y=trace2$y, xaxis=trace2$xaxis, yaxis=trace2$yaxis, hoveron=trace2$hoveron)
p <- add_trace(p, fill=trace3$fill, line=trace3$line, mode=trace3$mode, name=trace3$name, type=trace3$type, x=trace3$x, y=trace3$y, xaxis=trace3$xaxis, yaxis=trace3$yaxis, hoveron=trace3$hoveron)
p <- add_trace(p, line=trace4$line, mode=trace4$mode, name=trace4$name, type=trace4$type, x=trace4$x, y=trace4$y, xaxis=trace4$xaxis, yaxis=trace4$yaxis)
p <- layout(p, title=layout$title, xaxis=layout$xaxis, yaxis=layout$yaxis, margin=layout$margin)
p
 


```

Model Results SARIMAX { data-navmenu="Time Series Analysis"}
===================================================================

### Importance {.tabset}
-------------------------------------

### Delhi
```{r}
print("Delhi")
summary(ARIMAXD2.out)

```

### Mumbai
```{r}
print("Mumbai")
summary(ARIMAXM2.out)
```

### Chennai
```{r}

print("Chennai")
summary(ARIMAXC2.out)

```

Ljung - Box Residual Test { data-navmenu="Time Series Analysis"}
===================================================================

### Importance {.tabset}
-------------------------------------


### Delhi
```{r}

Box.test(ARIMAXD.out$residuals,lag=10,type="Ljung")

Box.test(x = ARIMAXD.out$residuals^2,lag=1,type="Ljung")


```

### Mumbai
```{r}

Box.test(ARIMAXM.out$residuals,lag=10,type="Ljung")

Box.test(x = ARIMAXM.out$residuals^2,lag=1,type="Ljung")


```

### Chennai
```{r}

Box.test(ARIMAXC.out$residuals,lag=10,type="Ljung")

Box.test(x = ARIMAXC.out$residuals^2,lag=1,type="Ljung")


```


Comparison & Results {data-navmenu="Time Series Analysis"}
===================================================================

Column
-------------------------------------

### Conclusion 
 Conclusion 
 In this project we have identified the SARIMA function using Hyndman & Khandakar algorithm in auto.arima to be the best model for our problem. Due to the limitations of computing power and relatively low quality of data, we were only able to forecast for Delhi and Mumbai. With distributed computing, this can be overcome. The model can be automated and scaled with better quality data to make an interactive dashboard for policy makers to make decisions and to track climate change and fluctuations. Technologies like HDFS, Map Reduce with R can be on a cloud service provider used for scaling. 


Thank You

### Delhi Final Prediction
```{r}

figD


```



Column
-------------------------------------

### Limitations 
 Limitations 
 The Ljung Box test confirmed conditional heteroscedasticity. GARCH models were fitted but the standard deviations were too high. The very nature of this data is highly volatile. Therefore we have obtained a reasonable prediction. Usage of a fourier curve for the seasonal component or modern techniques like LSTM Neural Nets might give better results.

### Mumbai Final Prediction
```{r}

figM


```

Web Scrapper 
=========================================

### Importance {.tabset}
-------------------------------------

### Page 1

```{r, eval=FALSE, echo=TRUE}
# This code was used to scrape meterological data from the web
data<-NULL
for ( year in 2019:2020)
{
  for(month in 1:12)
  {
    monthname<-c('january','february','march','april','may','june','july','august','september','october','november','december')
    if (monthname[month] %in% c('january','march','may','july','august','october','december')){j=31}
           else {j=30}
    if(month==2&year%%4==0){j=29}
    else if (month==2&year%%4!=0){j=28}
    for(day in 1:j)
    {
      url<-read_html(paste0('https://en.tutiempo.net/records/vidp/',day,'-',monthname[month],'-',year,'.html'))
temp.raw<-url%>%
  html_nodes('.Temp')%>%
  html_text()
temp.raw<-gsub('[^\x20-\x7E]', '', temp.raw)
humid.raw<-url%>%
  html_nodes('.hr')%>%
  html_text()
humid.raw<-str_remove(humid.raw,'%')
wind.raw<-url%>%
  html_nodes('.wind')%>%
  html_text()




   

```


### Page 2

```{r , eval=FALSE, echo=TRUE}

wind.raw<-str_remove(wind.raw,' km/h')
wind.raw<-na.omit(str_extract(wind.raw, '[[:digit:]]+'))
pressure.raw<-url%>%
  html_nodes('.prob')%>%
  html_text()

pressure.raw<-str_remove(pressure.raw,'hPa')
humidity<-mean(na.omit(as.integer(humid.raw)))
temp<-mean(na.omit(as.integer(temp.raw)))
pressure<-mean(na.omit(as.integer(pressure.raw)))
wind<-mean(na.omit(as.integer(wind.raw)))
Date<-as.Date(paste0(year,'-',month,'-',day))
datatemp<-data.frame(Date,temp,pressure,humidity,wind)
data<-rbind(data,datatemp)
    }  
    
}
  nametemp<-paste('data',year,sep='_')
  assign(nametemp,data)
  
}

data2<-rbind(data_2015,data_2016,data_2017,data_2018,data_2019,data_2020)

write.csv(data2,'data2.csv', row.names = FALSE)
```

Credits
=========================================

Column
-------------------------------------

### Group No
 Group No

28



### Member
 Wellington Daniel 

 2016IPM118


### Member
 Oshin Singhal 

 2016IPM070


Column
-------------------------------------
### Section
 Section 

A

### Member
 Shashank Giri

 2016IPM094


### Member
 Aditya Badonia 

 2016IPM004